High Gas Sensitivity to Nitrogen Dioxide of Nanocomposite ZnO-SnO2 Films Activated by a Surface Electric Field

Gas sensors based on the multi-sensor platform MSP 632, with thin nanocomposite films based on tin dioxide with a low content of zinc oxide (0.5–5 mol.%), were synthesized using a solid-phase low-temperature pyrolysis technique. The resulting gas-sensitive ZnO-SnO2 films were comprehensively studied by atomic force microscopy, Kelvin probe force microscopy, X-ray diffraction, scanning electron microscopy, transmission electron microscopy, scanning transmission electron microscopy, energy dispersive X-ray spectrometry, and X-ray photoelectron spectroscopy. The obtained films are up to 200 nm thick and consist of ZnO-SnO2 nanocomposites, with ZnO and SnO2 crystallite sizes of 4–30 nm. Measurements of ZnO-SnO2 films containing 0.5 mol.% ZnO showed the existence of large values of surface potential, up to 1800 mV, leading to the formation of a strong surface electric field with a strength of up to 2 × 107 V/cm. The presence of a strong surface electric field leads to the best gas-sensitive properties: the sensor’s responsivity is between two and nine times higher than that of sensors based on ZnO-SnO2 films of other compositions. A study of characteristics sensitive to NO2 (0.1–50 ppm) showed that gas sensors based on the ZnO-SnO2 film demonstrated a high sensitivity to NO2 with a concentration of 0.1 ppm at an operating temperature of 200 °C.

However, despite the good gas-sensitive properties of these materials, nanostructure synthesis technology is quite difficult to apply widely in the production of gas sensors. On the other hand, thin films of nanocomposite materials are used to determine low concentrations of gases (fractions and ppm units). Excellent results are shown by thin films based  COOH), and 1,4-dioxane as a solvent were used as precursors for the synthesis of SnO 2 -ZnO thin films. The reagents were purchased from "ECROS", Saint Petersburg, Russia. All chemicals used were of analytical grade or of the highest purity available.

Synthesis of ZnO-SnO 2 Thin Films
The synthesis was carried out in two stages. In the first stage, synthesis was carried out in a melt of abietic acid. The necessary amounts of zinc and tin salts were added to the melt, after which the melt was cooled and crushed. In the second stage, the necessary amounts of organic salts were dissolved in 1,4-dioxane so that the molar ratios of tin and zinc were Zn:Sn = 0:100 (sample 0ZnO), 0.5:99.5 (sample 0.5ZnO), 1:99 (sample 1ZnO), and 5:95 (sample 5ZnO) mol.%, respectively. The resulting solution was applied three times onto polycor substrates and a multi-sensor platform. The precursor solution was deposited on the multi-sensor platform using a mechanical pipette. Each layer was dried in air and in a drying cabinet at a temperature of 100 • C. Heat treatment was carried out at 500 • C for two hours ( Figure 1). The synthesis technique is described in more detail in previous studies [15,36]. To make gas sensors based on ZnO-SnO2 thin films, precursor was applied over a Pt counter-pin structure multi-sensor platform, MSP 632, and dried at 100 °C. Annealing at a temperature of 500 °C for two hours in air-and temperature-controlled conditions was carried out directly, using a heater and a temperature sensor MSP 632.
Electrophysical (volt-ampere characteristics; temperature dependences of electrical conductivity) and gas sensitivity measurements were estimated using an automated installation for determining the parameters of gas sensors [16]. In addition, the activation energy of conductivity (Ea) was estimated using the Arrhenius equation (Equation (1)) [48]: where k is the Boltzmann constant, and G0 the coefficient taking into account the bulk material conductivity. Next, the temperature-stimulated conduction measurements [16,49,50] method was used to evaluate the "effective" value of the energy barrier between the grains of a nanocrystalline material.
The gas response of sensors was measured with 5-50 ppm nitrogen dioxide (NO2) To make gas sensors based on ZnO-SnO 2 thin films, precursor was applied over a Pt counter-pin structure multi-sensor platform, MSP 632, and dried at 100 • C. Annealing at a temperature of 500 • C for two hours in air-and temperature-controlled conditions was carried out directly, using a heater and a temperature sensor MSP 632.
Electrophysical (volt-ampere characteristics; temperature dependences of electrical conductivity) and gas sensitivity measurements were estimated using an automated installation for determining the parameters of gas sensors [16]. In addition, the activation energy of conductivity (E a ) was estimated using the Arrhenius equation (Equation (1)) [48]: where k is the Boltzmann constant, and G 0 the coefficient taking into account the bulk material conductivity.
Next, the temperature-stimulated conduction measurements [16,49,50] method was used to evaluate the "effective" value of the energy barrier between the grains of a nanocrystalline material.
The gas response of sensors was measured with 5-50 ppm nitrogen dioxide (NO 2 ) balanced with synthetic air at operating temperatures of 200-250 • C. In the experiments, cylinders with test synthetic air and a mixture of synthetic air and NO 2 (Monitoring LLC, Saint Petersburg, Russia) were used. The gases were injected at a flow rate of 0.3 dm 3 /min using a gas mixture generator (Microgaz F, Moscow, Russia) [16]. The response of sensor elements based on ZnO-SnO 2 films was calculated using the following formula (Equation (2)): where R 0 and R g are the sensors resistance in synthetic air and in a mixture of synthetic air and NO 2, respectively. Subsequently, the gas sensor, which showed the best gas-sensitive properties, was tested for exposure to NO 2 with a concentration of 0.1 to 1 ppm.

Results
The XRD spectra of pure and ZnO-SnO 2 thin films are shown in Figure 2. All of the diffraction peaks are well indexed to the tetragonal structure of SnO 2 typical for cassiterite (JCPDS: 41-1445), regardless of the amount of modifying agents. Moreover, no additional diffraction peaks of ZnO were observed, which may be due both to the small amount of it in the system and to the small size of the crystallites. No other impurity phases are observed in the XRD pattern. The XRD patterns for pure and ZnO-SnO 2 materials show that the films have a polycrystalline microstructure.  The degree of material crystallinity was determined by XRD according to the method described in the paper [49,50]. It was found that an increase in the content of the introduced additives leads to an increase in the degree of materials' crystallinity (57.2, 66.0, 67.0 and 67.4% for 0ZnO, 0.5ZnO, 1ZnO, and 5ZnO materials, respectively).
Cross-sectional and top-view SEM images of 0ZnO (a), 0.5ZnO (c), 1ZnO (e), and 5ZnO (g) thin films are presented in . It was found that the films are cracked, homogenous, and uniform, with a continuous distribution of grains. The type of coating obtained is an area with adjacent grains, and the thickness of the three-layer films is 160-240 nm; statistical processing of the results of the SEM analysis is shown in the histograms below. However, in SEM images, as a rule, it is difficult to determine the dimensions of agglomeration of observed nanocrystallites. Therefore, the sizes of nanocrystallites measured may, in this case, be partly overestimated. The degree of material crystallinity was determined by XRD according to the method described in the paper [49,50]. It was found that an increase in the content of the introduced additives leads to an increase in the degree of materials' crystallinity (57.2, 66.0, 67.0 and 67.4% for 0ZnO, 0.5ZnO, 1ZnO, and 5ZnO materials, respectively).
Cross-sectional and top-view SEM images of 0ZnO (a), 0.5ZnO (c), 1ZnO (e), and 5ZnO (g) thin films are presented in Figure 3. It was found that the films are cracked, homogenous, and uniform, with a continuous distribution of grains. The type of coating obtained is an area with adjacent grains, and the thickness of the three-layer films is 160-240 nm; statistical processing of the results of the SEM analysis is shown in the histograms below. However, in SEM images, as a rule, it is difficult to determine the dimensions of agglomeration of observed nanocrystallites. Therefore, the sizes of nanocrystallites measured may, in this case, be partly overestimated. The 0.5ZnO thin film was investigated using the TEM method. The synthesized film material consists of many nanocrystallites with particles ranging in size from 4.5-11.5 nm (a,e−h). Where the film is thin, SnO2 nanocrystallites are visible in large quantities; between these, separate ZnO nanocrystallites are located. Studies conducted using the EDX method showed that the proportion of zinc in the material, by weight, does not exceed 1% of the mass (a-f). The mapping of the elements Zn, O, and Sn in one film's area is shown in b-d. It can be seen that the elements Sn, Zn, and O are evenly distributed throughout the material, without local agglomeration. The content of the elements Sn, Zn, and O is consistent with the relative amount of the elements Sn, Zn, and O in the initial solution.
The TEM images show that ZnO crystallites make contact with SnO2 crystallites. To prove this, high resolution photos were taken; areas 1 and 2 are highlighted in e,f. Area 2 comprises SnO2 nanocrystallites, which is confirmed by the EDX analysis. Area 1 consists of nanocrystallites of SnO2 and ZnO, which are uniformly distributed thoughout the area. Moreover, since there are significantly more SnO2 nanocrystallites, ZnO nanocrystallites make contact with SnO2 nanocrystallites. The 0.5ZnO thin film was investigated using the TEM method. The synthesized film material consists of many nanocrystallites with particles ranging in size from 4.5-11.5 nm (Figure 4a,e-h). Where the film is thin, SnO 2 nanocrystallites are visible in large quantities; between these, separate ZnO nanocrystallites are located. Studies conducted using the EDX method showed that the proportion of zinc in the material, by weight, does not exceed 1% of the mass (Figure 4a-f). The mapping of the elements Zn, O, and Sn in one film's area is shown in Figure 4b-d. It can be seen that the elements Sn, Zn, and O are evenly distributed throughout the material, without local agglomeration. The content of the elements Sn, Zn, and O is consistent with the relative amount of the elements Sn, Zn, and O in the initial solution. The TEM images show that ZnO crystallites make contact with SnO 2 crystallites. To prove this, high resolution photos were taken; areas 1 and 2 are highlighted in Figure 4e,f. Area 2 comprises SnO 2 nanocrystallites, which is confirmed by the EDX analysis. Area 1 consists of nanocrystallites of SnO 2 and ZnO, which are uniformly distributed thoughout the area. Moreover, since there are significantly more SnO 2 nanocrystallites, ZnO nanocrystallites make contact with SnO 2 nanocrystallites.  To observe the planes of the crystal structure, images with high magnification were obtained for some sections of the film. One particle with a lattice fringe of 0.264 nm corresponds to the plane (002) of the ZnO hexagonal structure: it is shown in h. Additionally, on this site there are areas in which it is possible to distinguish a lattice fringe of 0.334 nm, which corresponds to the plane (110) of the SnO2 rutile structure. Thus, it can be concluded that nanoparticles consist of ZnO and SnO2 crystals.  To observe the planes of the crystal structure, images with high magnification were obtained for some sections of the film. One particle with a lattice fringe of 0.264 nm corresponds to the plane (002) of the ZnO hexagonal structure: it is shown in Figure 4h AFM and KPFM images of 0ZnO, 0.5ZnO, 1ZnO, and 5ZnO films are presented in Figure 5. AFM studies showed that the films have a granular structure, with a height difference of Sq of 10-35 nm (). In general, with a decrease in the content of ZnO in the film, the roughness decreases from 34 to 11 nm. However, the 0.5ZnO film surface has a higher roughness than the surface of the 0ZnO and 1ZnO films-see (curve 2). AFM studies showed that the films have a granular structure, with a height difference of Sq of 10-35 nm (Figures 5 and 6). In general, with a decrease in the content of ZnO in the film, the roughness decreases from 34 to 11 nm. However, the 0.5ZnO film surface has a higher roughness than the surface of the 0ZnO and 1ZnO films-see Figure 6 (curve 2). KPFM studies have shown that a lowest average value of the surface potential ( (curve 1)) of about 4 mV is characteristic of SnO2 film (h). 1ZnO and 5ZnO films have similar average values of surface potential, equal to 15-30 mV (b,d).
However, for 0.5ZnO film, a peak value of the surface potential is observed, the average value of which reaches 845 mV, and at some points on the surface the maximum value is 1824 mV; the difference between the maximum and minimum values of Vb could be 1642 mV (Figures 5f and S2).
The samples' surface composition () was also calculated from high-resolution XPS spectra using relative sensitivity factors from the Scofield Library. The high-resolution spectra of Zn2p, Sn3d and the valence band are presented in . It can be seen from b that the photoelectronic line Sn 3d for the sample 0ZnO (red curve) corresponds to SnO2.   However, for 0.5ZnO film, a peak value of the surface potential is observed, the average value of which reaches 845 mV, and at some points on the surface the maximum value is 1824 mV; the difference between the maximum and minimum values of V b could be 1642 mV (Figure 5f and Figure S2).
The samples' surface composition (Table 2) was also calculated from high-resolution XPS spectra using relative sensitivity factors from the Scofield Library. The high-resolution spectra of Zn2p, Sn3d and the valence band are presented in Figure 7. It can be seen from Figure 7b that the photoelectronic line Sn 3d for the sample 0ZnO (red curve) corresponds to SnO 2 . KPFM studies have shown that a lowest average value of the surface potential ( (curve 1)) of about 4 mV is characteristic of SnO2 film (h). 1ZnO and 5ZnO films have similar average values of surface potential, equal to 15-30 mV (b,d).
However, for 0.5ZnO film, a peak value of the surface potential is observed, the average value of which reaches 845 mV, and at some points on the surface the maximum value is 1824 mV; the difference between the maximum and minimum values of Vb could be 1642 mV (Figures 5f and S2).
The samples' surface composition () was also calculated from high-resolution XPS spectra using relative sensitivity factors from the Scofield Library. The high-resolution spectra of Zn2p, Sn3d and the valence band are presented in . It can be seen from b that the photoelectronic line Sn 3d for the sample 0ZnO (red curve) corresponds to SnO2.    Minor changes in the maximum positions of the Sn 3d photoelectronic lines for samples containing ZnO suggest the formation of a special interface at the nanocomposite oxides' surface, which leads to a decrease in the work function. Analysis of the XPS spectra valence band edge makes it possible to evaluate and determine in more detail both the change in the maximum of the valence band (VBM), and the band gap and the Fermi level position at the phase boundary [51][52][53][54]. The formation of this interface in the ZnO-SnO 2 system can be confirmed by analysing the high-resolution Zn2p spectra (Figure 7a). At zinc concentrations of 0.5% and 1%, the shapes of the Zn2p photoelectronic lines are very different from the case of bulk ZnO ( Figure S1). In the region of binding energies of 1033.5 eV, plasmon losses, which are inherent in metallic zinc, are visible; however, the position and shape of the Zn2p3 peak exclude the presence of metallic zinc. The ratio of peak intensities of Zn2p3 photoelectronic lines' plasmon losses for the obtained samples, as well as a comparison for pure metallic and ZnO crystalline, is shown in Figure 8a. Thus, the presence of a gap between the Fermi level and the edge of the valence band on the one hand, and the features of Zn2p photoelectronic lines on the other, completely exclude the presence of metallic zinc in the synthesized nanocomposites. The peculiarity of the Zn2p photoelectronic lines here can be linked to the formation of an interface enriched with charge carriers at the phase boundary, which leads to inelastic scattering of photoelectrons on them, and the appearance of an intense peak of plasmon losses in the region of 1033.5 eV. Minor changes in the maximum positions of the Sn 3d photoelectronic lines for samples containing ZnO suggest the formation of a special interface at the nanocomposite oxides' surface, which leads to a decrease in the work function. Analysis of the XPS spectra valence band edge makes it possible to evaluate and determine in more detail both the change in the maximum of the valence band (VBM), and the band gap and the Fermi level position at the phase boundary [51][52][53][54]. The formation of this interface in the ZnO-SnO2 system can be confirmed by analysing the high-resolution Zn2p spectra (a). At zinc concentrations of 0.5% and 1%, the shapes of the Zn2p photoelectronic lines are very different from the case of bulk ZnO ( Figure S1). In the region of binding energies of 1033.5 eV, plasmon losses, which are inherent in metallic zinc, are visible; however, the position and shape of the Zn2p3 peak exclude the presence of metallic zinc. The ratio of peak intensities of Zn2p3 photoelectronic lines' plasmon losses for the obtained samples, as well as a comparison for pure metallic and ZnO crystalline, is shown in a. Thus, the presence of a gap between the Fermi level and the edge of the valence band on the one hand, and the features of Zn2p photoelectronic lines on the other, completely exclude the presence of metallic zinc in the synthesized nanocomposites. The peculiarity of the Zn2p photoelectronic lines here can be linked to the formation of an interface enriched with charge carriers at the phase boundary, which leads to inelastic scattering of photoelectrons on them, and the appearance of an intense peak of plasmon losses in the region of 1033.5 eV.

Materials
The values of the valence band edge from high-resolution XPS spectra close to the Fermi level are determined using a linear approximation of the background line and the useful signal. The intersection of the obtained lines determines the value of the valence band edge, shown in c.
where ∆E CL = E Zn2p ZnO − E Sn3d SnO 2 ; The values of the valence band edge from high-resolution XPS spectra close to the Fermi level are determined using a linear approximation of the background line and the useful signal. The intersection of the obtained lines determines the value of the valence band edge, shown in Figure 7c.
From the analysis of the valence band edge, the values of the change in the valence band maximum for each sample are obtained according to the well-known formula (Equation (3)) [55]:  Table 3. The distribution of the change in the valence band level at the interface for each sample is presented below in Figure 8b.
The dependence of reverse resistance on reverse temperature for all samples has a characteristic form for metal oxide semiconductors (Figure 9). Valence band edge values and the binding energy of core levels for ZnO2p3/2 and SnO2 3d5/2 are presented in . The distribution of the change in the valence band level at the interface for each sample is presented below in b.
The dependence of reverse resistance on reverse temperature for all samples has a characteristic form for metal oxide semiconductors (). The results of the gas sensitivity study of ZnO-SnO2 films at 200 °C and 250 °C are shown in . The typical response of gas sensors when exposed to NO2 with concentrations of 5, 10, and 50 ppm is shown in c. The activation energy of conductivity (E a ), calculated by the Arrhenius equation (Figure 9b, curve 1), shows a sharp (by 2-2.5 times) decrease in the E a and the potential barrier Φ B for 0.5ZnO films. The E a decreases from 0.51 eV to 0.2 eV, and the Φ B from 1.1 eV to 0.54 eV.
The results of the gas sensitivity study of ZnO-SnO 2 films at 200 • C and 250 • C are shown in Figure 10. The typical response of gas sensors when exposed to NO 2 with concentrations of 5, 10, and 50 ppm is shown in Figure 10c. It is shown (c) that, when exposed to NO2 concentrations of 5, 10, and 50 ppm, a gas sensor based on a 0.5ZnO-99.5SnO2 film is approximately 1.5-9 times more sensitive to NO2 compared to gas sensors based on ZnO-SnO2 films with other compositions. For exposure to NO2 with a concentration of 50 ppm and an exposure temperature of 200 °C, the sensor's response S is equal to 146. The evaluation of the response time tresp shows () that for this sensor, when exposed to 5 ppm NO2, tresp is 144 s and 80 s at operating temperatures of 200 and 250 °C, respectively. A gas sensor based on a 0.5ZnO-99.5SnO2 film has a higher sensitivity; it was therefore investigated for exposure to NO2 with concentrations from 0.1 to 1 ppm. The sensor response of this gas sensor is shown in d. The response value, when exposed to NO2 with It is shown (Figure 10c) that, when exposed to NO 2 concentrations of 5, 10, and 50 ppm, a gas sensor based on a 0.5ZnO-99.5SnO 2 film is approximately 1.5-9 times more sensitive to NO 2 compared to gas sensors based on ZnO-SnO 2 films with other compositions. For exposure to NO 2 with a concentration of 50 ppm and an exposure temperature of 200 • C, the sensor's response S is equal to 146. The evaluation of the response time t resp shows ( Table 4) that for this sensor, when exposed to 5 ppm NO 2 , t resp is 144 s and 80 s at operating temperatures of 200 and 250 • C, respectively. A gas sensor based on a 0.5ZnO-99.5SnO 2 film has a higher sensitivity; it was therefore investigated for exposure to NO 2 with concentrations from 0.1 to 1 ppm. The sensor response of this gas sensor is shown in Figure 10d. The response value, when exposed to NO 2 with a concentration of 1.0 ppm, is 3.9, and, when exposed to NO 2 with a concentration of 0.1 ppm, is 1.1.
Thus, gas sensors based on a ZnO-SnO 2 film, synthesized using a solid-phase pyrolysis technique with a ratio of 0.5ZnO, have a range of measured concentrations of NO 2 equal to 0.1-50 ppm, with a response time of 80-144 s, and an operating temperature of 200 • C.
To test the selectivity of the sensor based on a 0.5ZnO-99.5SnO 2 film to other gases, studies were conducted with acetone, ethanol, ammonia, and water molecules, with a concentration of 200 ppm at an operating temperature of 200 • C. The sensor response was 1.1 for acetone, 1.25 for ethanol, 1.76 for ammonia, and 1.1 for water. It can be seen that, in comparison with the response value for nitrogen dioxide with a concentration of 50 ppm (146), these values do not exceed 1.2%.

Discussion
It is known that the valence band edge for pure SnO 2 is 3.5 eV [54,55]. This is also confirmed by our measurements for the material ZnO:SnO 2 = 0:100 ( Figure 8, Table 3). For 0.5ZnO film materials or for other ZnO-SnO 2 materials, the energy level of the valence band edge decreases to 2.95 eV. Further decrease in the valence band edge value is slower-up to 2.77 and 2.74 eV for 1ZnO and 5ZnO materials, respectively. Thus, the energy of the valence band edge in ZnO-SnO 2 films decreases with an increase in the ZnO concentration.
Potential barriers are formed at the interface of SnO 2 -SnO 2 homostructures and the ZnO-SnO 2 heterojunction. Measured by temperature-stimulated conduction measurements, the value of the potential barrier Φ B for the SnO 2 film was 0.80 eV (Figure 9b). This potential barrier is formed due to the adsorption of oxygen molecules on the surface of SnO 2 nanocrystals, and their subsequent ionization with the formation of O 2 − ions [33,[56][57][58][59][60]. The contact of two SnO 2 nanocrystallites is shown in Figure 11a. It is well known that, because of this contact, a depletion region is formed in SnO 2 nanocrystallites [59][60][61]. As can be seen from , gas sensors based on the film material 0.5ZnO-99.5SnO2 show a peak response to NO2 exposure, compared to other samples.
The reason for this sharp increase in response may, on the one hand, be a decrease in the potential barrier ΦB [65]. On the other hand, however, the reason for the sharp increase in the response of gas sensors based on film material 0.5ZnO-99.5SnO2 is, in our opinion, the high surface potential formed due to the contact of ZnO and SnO2 crystallites. The analysis of TEM images () shows that the distance between the crystallites is no more than 0.1 nm. At such values, distances, and potentials, there is a surface electric field with a strength of 2 × 10 7 V/cm. It is known that an electric field with a strength of up to 3 × 10 7 V/cm can exist in oxide materials if their thickness is up to 4 nm [66,67]. The presence of a surface electric field is supported by the response decreasing from 200 °C to 250 °C (a,b), since it is well known that an increase in temperature leads to a decrease in the influence of the electric field. The effect of the electric field influence on the atoms' diffusion over the surface, and in the volume of solids, has been well studied [68,69].
As shown in [70], the surface electric field of this magnitude, created by charged adsorption centers, significantly affects the mechanism of the interaction of polar gas molecules with the gas-sensitive material's surface.
It is known that the NO2 molecule has a dipole moment 0.29-0.33 D [71][72][73]. In this case, the interaction energy of the polar adsorbate molecule with the charged adsorption center (Qaa) can be estimated using the following formula (Equation (4)) [74]: where e is the electron charge; V is the valence of the adsorbent ion; μ is the dipole moment of the adsorbate molecule; zо is the equilibrium distance between the adsorbate molecule The study of the gas sensor based on the film material 0.5ZnO-99.5SnO 2 showed that the value of Φ B sharply decreases to 0.47 eV (Figure 9b). With a further increase in the ZnO concentration (1ZnO and 5ZnO), Φ B becomes significantly higher-1.1 and 1.0 eV, respectively. The mechanism of a sharp decrease in Φ B in the gas sensor based on the 0.5ZnO-99.5SnO 2 film is shown in Figure 11b. It is known that the work function of ZnO (5.2-5.3 eV) [62,63] is higher than that of SnO 2 (4.8-4.9 eV) [61,64]. In addition, the Fermi level of SnO 2 is closer to the vacuum level, in relation to the Fermi level of ZnO. When ZnO and SnO 2 crystallites are in contact with each other, electrons will transfer from tin dioxide to zinc oxide ( Figure 11). The value of the potential barrier at the SnO 2 -ZnO-SnO 2 heterojunction will decrease in this case (Figure 9b). According to TEM studies (Figure 4c), individual ZnO crystallites come into contact with SnO 2 crystallites, so ZnO becomes highly enriched with charge carriers. This fact is also confirmed by XPS studies: plasmon losses are observed on Zn2p3 zinc photoelectronic lines (Figure 7a). The same effect leads to the appearance of regions with large local surface potential Vb. This is confirmed by our KPFM measurements (Figures 5c and 6b). On the surface of the 0.5ZnO film, the average value of the surface potential is 845 mV, and in some places reaches values of 1824 mV (Figure 5c). For film materials with a higher content of zinc oxides (1ZnO and 5ZnO), the concentration of ZnO crystallites becomes higher, which leads to a decrease in the average value of the surface potential to 15-30 mV. The existence of high values of the V b in 0.5ZnO-99.5SnO 2 films leads to the formation of a strong surface electric field. The electric field contributes to the reduction of the Φ B between nanocrystallites. In this regard, E a also decreases in the 0.5ZnO-99.5SnO 2 film. In addition, it can be noted that the surface potential affects the morphology of the film's surface ( Figure 6).
As can be seen from Figure 10, gas sensors based on the film material 0.5ZnO-99.5SnO 2 show a peak response to NO 2 exposure, compared to other samples.
The reason for this sharp increase in response may, on the one hand, be a decrease in the potential barrier Φ B [65]. On the other hand, however, the reason for the sharp increase in the response of gas sensors based on film material 0.5ZnO-99.5SnO 2 is, in our opinion, the high surface potential formed due to the contact of ZnO and SnO 2 crystallites. The analysis of TEM images (Figure 4) shows that the distance between the crystallites is no more than 0.1 nm. At such values, distances, and potentials, there is a surface electric field with a strength of 2 × 10 7 V/cm. It is known that an electric field with a strength of up to 3 × 10 7 V/cm can exist in oxide materials if their thickness is up to 4 nm [66,67]. The presence of a surface electric field is supported by the response decreasing from 200 • C to 250 • C (Figure 10a,b), since it is well known that an increase in temperature leads to a decrease in the influence of the electric field. The effect of the electric field influence on the atoms' diffusion over the surface, and in the volume of solids, has been well studied [68,69].
As shown in [70], the surface electric field of this magnitude, created by charged adsorption centers, significantly affects the mechanism of the interaction of polar gas molecules with the gas-sensitive material's surface.
It is known that the NO 2 molecule has a dipole moment 0.29-0.33 D [71][72][73]. In this case, the interaction energy of the polar adsorbate molecule with the charged adsorption center (Q aa ) can be estimated using the following formula (Equation (4)) [74]: where e is the electron charge; V is the valence of the adsorbent ion; µ is the dipole moment of the adsorbate molecule; z 0 is the equilibrium distance between the adsorbate molecule and the charged adsorption centre; and Na is the Avogadro constant. Calculations show that if z 0 is equal to 0.2 nm, the Q aa for the NO 2 molecule is 31-35 kJ/mol. At the same time, we have shown that a sufficiently strong surface electric field (Es), of magnitude (1-2) × 10 7 V/cm, is created at the crystallite boundary of tin and zinc oxides. The processes of adsorption and dissociation of NO 2 molecules are activated by a strong surface electric field. NO 2 molecules receive an additional field-induced dipole moment (µ i ), the magnitude of which depends on the polarizability of the α molecule. It is known that the NO 2 molecule has a high polarizability (α), equal to 1.8 × 10 −24 cm 3 [67].
When the amount of polarizability increases, the surface electric field effect on the molecule also increases. The energy of the adsorbate-adsorbent interaction (Q aa ) increases by the value of the induction component (Q i ) associated with the occurrence of the induced dipole moment. The value of µ i can be estimated using the formula (Equation (5)) [75].
Calculation of the total dipole moment (µ = µ + µ i ) shows that µ can be equal to 0.47-0.51 D.
In addition, the adsorbate molecule will be affected by orientation polarization, in accordance with the direction of the electric field, which depends on the magnitude of the molecule's dipole moments [75]. The energy of the orientation influence (Q or ) of the field is equal to (Equation (6)): The total energy of the adsorbate adsorbent (Q aa ) interaction between the adsorption centre and the NO 2 molecule is calculated by the formula (Equation (7)): Q aa = (Q aa + Q i + Q or ).
The calculation results show that, for an NO 2 molecule with high polarizability, the Q aa can increase to 57-60 kJ/mol, that is, almost twofold. Such energy may already be enough to initiate the process of dissociation of gas molecules, since the activation energy of the molecule's dissociation is half of the molecule dissociation energy [76,77]. The latter circumstance means that dissociation of the NO 2 molecule is possible. These arguments do not contradict previously published works [78,79]. Thus, adsorption of NO 2 molecules, with their subsequent dissociation, can occur (Equation (8)): A strong surface electric field is a powerful activating factor, as is ultraviolet radiation. In the presence of these factors, forced ionization of adsorbed molecules can occur (Equation (9)) [80]: As can be seen, these reactions can lead to a stronger response to NO 2 molecules. A comparative analysis of the gas sensitivity to NO 2 of nanocomposite ZnO-SnO 2 structures presented in Table 1 and in this paper shows that the size of nanostructures (nanocrystallites) is essential for high gas sensitivity to low concentrations of NO 2 in the air. In all works, the nanostructures of the gas-sensitive material are formed by different technologies, and have sizes from 5 to 17 nm [15,29,30,[40][41][42]. In addition, when the operating temperature of the gas sensor decreases, the response and recovery times increases [30,81]. The optimal value of the operating temperatures for sensors with low concentrations of nitrogen dioxide is, in our opinion, in the range of 150-250 • C, in order to achieve the optimal value of response times of 80-140 s. When analyzing these articles, the gas-sensitive properties of sensors based on ZnO-SnO 2 films are similar.
The characteristics of NO 2 sensors based on other materials are presented in Table 5. Sensors operating at room temperature have long response/recovery times [82][83][84][89][90][91][92][93]. The advantage of such sensors is that there is no need for heating to operating temperature, and they can be used indoors. Since gas sensors are used for NO 2 monitoring outside, in a wide temperature range, sensors operating at 200 • C are recommended for analysis unification. Sensors with response/recovery times in the range of 30-180 s have a more complex technology for forming a gas-sensitive layer [86][87][88]93]. Some gas-sensitive materials are unstable over time, in particular metal sulfides [83,85,86].
Thus, it can be seen that many materials used in gas sensors of the resistive type have similar gas-sensitive characteristics. The choice of gas-sensitive materials is determined by the following requirement: "simple technology-high-quality, reproducible and fast response to a given concentration." The solid-phase pyrolysis method proposed in this work and in our other works [15,36] satisfies this requirement. Using the described method, it is possible to obtain composite nanoscale materials of various compositions with high values of gas-sensitive characteristics. In this case, by selecting the composition of mixed ZnO-SnO 2 oxides, this technology allows for the formation of heterostructures with a strong surface electric field, thereby "fine-tuning" the gas-sensitive properties.

Conclusions
Nitrogen dioxide gas sensors were formed on the MSP 632 multi-sensor platform. The gas-sensitive element of the sensor is a ZnO-SnO 2 film with thickness 60-200 nm. The film materials were formed using a new low-temperature pyrolysis technique; the zinc oxide content was from 0.5 to 5 mol.%.
Comprehensive studies of the materials' physicochemical, structural, electrophysical, and other properties were carried out using modern research methods. It was shown that all ZnO-SnO 2 films are composite with crystallite sizes of 4-30 nm and have a cassiterite structure. Zinc oxide nanocrystallites are evenly distributed in the structure of ZnO-SnO 2 films. This leads to the creation, in the 0.5ZnO-99.5SnO 2 film, of a surface potential of up to 1800 mV. This is because of the existence of a strong surface electric field with a strength of (1-2) × 10 7 V/cm. When NO 2 molecules approach the surface of the ZnO-SnO 2 film, a strong surface electric field promotes the induction of a dipole moment in the molecule. As a result, this leads to an increase in the energy of the adsorbate-adsorbent interaction between the adsorption center and the NO 2 molecule, and an activation of the molecules' dissociation process. The response of the gas sensor based on 0.5ZnO film is 1.5-9 times higher compared to sensors based on other ZnO-SnO 2 films.
Thus, it was shown that the gas sensor based on film material 0.5ZnO-99.5SnO 2 is very promising for the detection of NO 2 gas at concentrations of 0.1-50 ppm; it has a response time equal to 80-144 s and an operating temperature of 200 • C.